FTI Consulting News Bytes – 14 November 2025

FTI Consulting News Bytes

This week, reports emerged that Meta Chief AI Scientist Yann LeCun plans to leave the company and build his own startup to continue his work on ‘world models’. In more AI-related news, Anthropic announced plans to spend $50bn to build custom data centres in several US locations, including Texas and New York, in partnership with London-based Fluidstack. It was a significant week for regulation, as the UK government unveiled plans to give regulators greater powers to impose fines on companies that fail to comply with cybersecurity requirements. Elsewhere, the BBC’s editorial standards and guidelines came under scrutiny following an internal memo by Michael Prescott who highlighted a series of failures at the broadcaster, including coverage of Trump. And finally, IBM revealed two new supercomputers called Loon and Nighthawk – contenders in the race to build an error-free quantum computer. 

This week’s news

Meta AI pioneer to launch own startup

Meta’s Chief Artificial Intelligence Scientist Yann LeCun is planning to leave the company to launch his own start-up, as Mark Zuckerberg seeks to overhaul the company’s AI operations. LeCun, a Turing Award winner who is considered one of the trailblazers of modern AI, is reportedly leaving Meta in the coming months and is already in talks to raise capital for a startup focused on advancing his work on world models. According to TechCrunch, LeCun’s departure comes at a critical time for Meta, which has changed how it approaches AI development and has started revamping its AI organisation after hiring over 50 engineers and researchers from competitors to build out its Meta Superintelligence Labs.

Anthropic to invest $50bn in US data centres

Anthropic plans to invest $50bn in building artificial intelligence infrastructure in the US over the coming years, as the company races to secure new computing power. It said it would develop new data centers in New York and Texas in partnership with Fluidstack, “with more sites to come.” The first locations are slated to go live in 2025, creating 800 permanent jobs and more than 2,000 construction roles. CNBC noted that the investment positions Anthropic as a major domestic player in physical AI infrastructure at a moment when policymakers are increasingly focused on US based compute capacity and technological sovereignty. 

UK regulators set to crack down harder on cyber security failures

The UK government plans to grant regulators expanded powers to fine companies for failing to comply with cyber security rules, as ministers estimate that attacks now cost the UK almost £15bn each year. The Financial Times reports that the Cyber Security and Resilience Bill includes measures that would give the sector regulators the power to fine certain companies up to 4% of their annual turnover, or £17mn if that is larger, if they fail to comply with the rules. Under the proposals, medium and large IT services companies providing IT management, help desk support and cybersecurity to critical services face regulation for the first time and will be required to report potentially significant cyber security breaches to regulators and the National Cyber Security Centre.

Internal memo sparks scrutiny of BBC editorial standards

The BBC’s editorial standards and guidelines have come under scrutiny following an internal memo by Michael Prescott, a former advisor to its standards committee, who accused the broadcaster of a series of failures in its coverage of Trump, the Gaza war and the rights of transgender people in September. According to the Financial Times, Prescott singled out a BBC Panorama programme broadcast in October 2024, which allegedly spliced together clips from separate parts of an inflammatory speech Trump made in Washington on January 6, 2021 – the day his supporters stormed Congress as lawmakers ratified his loss to Joe Biden in the 2020 US presidential. In the programme, Trump was shown telling his supporters that “we’re going to walk down to the Capitol” and that they would “fight like hell,” though the remarks were originally made at different points in his speech. 

IBM has unveiled two complex quantum computers

IBM has announced the launch of two new supercomputers, named Loon and Nighthawk, aimed at achieving error-free computing and substantiating existing research. According to Reuters, Loon is still in the early stages and IBM has not disclosed when external users will be able to test it. The company believes that Nighthawk could beat some classical computers on some tasks by the end of next year and is collaborating with a group of startups and researchers to share its code openly so that others can test and use it. New Scientist quoted Stephen Bartlett at the University of Sydney in Australia, who said that while more testing and benchmarking of the new devices is necessary, the increased qubit connectivity is exciting.

Top Tweets of the Week

• Greg Bensinger, Tech Reporter at Reuters: “An important and timely question: what recourse is there when generative AI systems defame? A new set of lawsuits could set an important precedent, by @kenbensinger.”
• Sam Altman, CEO at OpenAI: “GPT-5.1 is out! It’s a nice upgrade. I particularly like the improvements in instruction following, and the adaptive thinking. The intelligence and style improvements are good too.”
• Bloomberg: “Microsoft plans to use access to OpenAI’s development of custom AI semiconductors to help its own chip effort.”

Number of the week

 £3.6bn The amount pulled from equity funds in four months ending in October, according to Calastone, the largest outflow ever recorded.